I'm commenting on a single aspect of the proposed patch
here, not about the bug itself. Please also note this
hunk isn't needed anymore, since I removed the affected
lines entirely.
On Wed, 2 Oct 2024 20:43:26 +0200 Helmut Grohne <hel...@subdivi.de> wrote:
..
@@ -133,11 +168,13 @@
-e "s$(xSEDx) $${o}=[^=]+=\.$(xSEDx)$(xSEDx)g;" \
obj-$(DEB_HOST_GNU_TYPE)/winpr/include/winpr/buildflags.h
+ifeq ($(filter nocheck,$(DEB_BUILD_OPTIONS)),)
execute_after_dh_auto_build:
dh_auto_build --builddirectory=debian/testtmp
execute_after_dh_auto_test:
dh_auto_test --builddirectory=debian/testtmp
+endif
This seems to be wrong way to do it, because debhelper doesn't
implement full make(1) machinery. If we're to do this, it should
be done like:
execute_after_dh_auto_build:
+ifeq ($(filter nocheck,$(DEB_BUILD_OPTIONS)),)
dh_auto_build --builddirectory=debian/testtmp
+endif
execute_after_dh_auto_test:
+ifeq ($(filter nocheck,$(DEB_BUILD_OPTIONS)),)
dh_auto_test --builddirectory=debian/testtmp
+endif
In other words: keep the targets, but make them empty
based on condition. Instead of conditionally hide the
targets, which confuses debhelper.
Thanks,
/mjt
